Heroku Outage History

A disruption affected Cedar Private Spaces apps hosted in Virginia, where users experienced failures when starting and scaling dynos. The disruption began at 1:41 PM UTC as Heroku noted engineers were investigating.

At 2:01 PM UTC, engineers identified an upstream service provider issue that was impacting Cedar Private Spaces in Virginia. They were working with the upstream provider to expedite a resolution. A follow-up at 2:16 PM UTC indicated some service recovery had been observed and that teams were monitoring for full resolution.

Finally, the status was updated to indicate recovery as services returned to normal by 2:36 PM UTC. The impact appeared regional to Virginia, with no widespread outages reported.

The incident impacted Heroku's Apps and Data components, likely causing interruptions for users attempting to run apps or access data. The outage began at Feb 18, 2026 1:51 AM UTC and was resolved by Feb 18, 2026 3:11 AM UTC, lasting about 1 hour and 20 minutes.

Heroku's status page documented two updates during the event. The initial update at Feb 18, 2026 1:51 AM UTC noted a service disruption and that engineers were investigating. A follow-up update at Feb 18, 2026 3:11 AM UTC indicated recovery, with the Apps and Data components returning to normal. The incident involved two components and was resolved without extended regional impact.

User reports were limited, reflecting a relatively contained issue. Overall, the outage status progressed from a disruption to normal service within the 1 hour 20 minute window.

The Heroku Tools component experienced a feature disruption that caused API access and dashboard/deployments issues for users. The outage lasted about 1 hour and 15 minutes, beginning when the status page moved from up to warn at Jan 15, 2026 12:46 PM UTC and ending when it returned to up at Jan 15, 2026 2:01 PM UTC. The incident was publicly described as a feature disruption affecting the Tools component during investigations by Heroku engineers.

Heroku published two status updates during the incident to communicate investigation progress and restoration. User reports came from multiple regions, including Brazil, Kenya, Pakistan, Poland, France, and the United States, indicating broadly distributed impact on APIs, deployments, and dashboard access. The Tools component subsequently returned to normal operation as the issue was resolved.

StatusGator detected login and CLI access issues for Heroku at Jan 15, 2026 12:25 PM UTC, about 21 minutes before the official acknowledgement at 12:46 PM UTC. The early warning ended at 1:25 PM UTC. The disruption affected the Tools component and impacted users globally, hindering login, API use, CLI commands, and dashboard access. The first official update appeared at 12:46 PM UTC indicating engineers were investigating and that the Tools component had degraded. The incident was resolved by 2:01 PM UTC when the Tools component returned to up, signaling recovery. The provider issued 2 status updates during the incident.

An outage affected the Heroku platform and multiple Salesforce cloud services, including Commerce Cloud, Core services, Hyperforce, MuleSoft, Marketing Cloud, Slack, and Tableau. The disruption impacted app hosting, data processing, and integrations for many customers. The outage began on Oct 20, 2025 8:49 AM UTC and was resolved by Oct 21, 2025 4:16 AM UTC, lasting about 19 hours and 27 minutes. Affected components included Tools, Data, and Apps.

Salesforce's status updates indicated the issue was triggered by a third-party vendor, with engineers working to resolve the disruption. The provider issued three official updates during the incident before restoration, noting ongoing remediation and eventual recovery.

Heroku reported an outage impacting the Apps component of the platform due to elevated H99 errors in the runtime, affecting users in the US and EU. The issue began at 11:17 PM UTC and resulted in applications failing to start or respond in those regions.

Status updates documented the incident: The first update at Sep 25, 2025 1:11 AM UTC indicated investigation into elevated H99 errors. A subsequent update at Sep 25, 2025 2:16 AM UTC showed the Apps component had recovered, returning to a normal state. The outage lasted roughly 2 hours and 59 minutes.

The Heroku Apps component experienced an outage starting at Sep 24, 2025 11:40 PM UTC with elevated H99 errors affecting US and EU regions, causing downtime for user applications. The Apps component was later reported down from Sep 25, 2025 12:11 AM UTC and was restored by Sep 25, 2025 1:01 AM UTC as engineers worked to resolve the issue. The outage lasted about 1 hour and 21 minutes and affected users across two regions.

The Heroku Data component in the US region experienced a brief degradation as API responses from the gpt-oss-120b model declined. Users may have encountered slower or failing API calls during the incident window. The outage began at Sep 22, 2025 4:26 AM UTC and was resolved by Sep 22, 2025 1:46 PM UTC, lasting approximately 9 hours and 20 minutes.

Heroku issued 2 status updates during the incident, first noting an investigation into the issue and later confirming restoration of normal service. The impact was focused on the US region and affected the Data component.

The incident affected Heroku DNS for New Apps and Domains, making DNS names for new apps and domains unreachable and causing delays provisioning DNS records. The issue began on Sep 16, 2025 4:21 PM UTC and continued until Sep 17, 2025 11:31 AM UTC, lasting about 19 hours and 10 minutes. Users attempting to provision new apps or add custom domains were impacted, with delays centered on DNS record provisioning. The Apps component was listed with a degraded status during the event, and the provider issued two status updates—one announcing an investigation and a later one noting restoration of service.

Heroku experienced degraded performance across the Cedar Common Runtime due to a kernel update from an upstream provider, which caused increased dyno relocations. Apps could see logs stating 'Relocating dyno to a new server', potentially affecting runtime stability and deployments. Separately, Heroku Connect Add-on provisioning had issues in the US and EU, with a workaround to cancel during step 2 and refresh the page.

Timeline and duration: The Apps component moved to a degraded state at Sep 9, 2025 10:56 PM UTC. Tools progressed to up by Sep 10, 2025 9:21 PM UTC, while Apps remained affected until Sep 11, 2025 11:46 PM UTC, for a total duration of roughly 54 hours. The provider issued 4 official status updates during the incident.

Geographic and user impact: The Connect Add-on issues were region-specific to the US and EU, while Cedar Common Runtime dyno relocations affected apps more broadly. All components were restored by Sep 11, 2025 11:46 PM UTC. Users may have experienced intermittent disruptions or deployment delays during the degraded period.